Today on the Datanauts we’re talking about in-memory databases. The idea is that you can load up a host with lots of RAM, cram your database in there, and get more transactions due to the lower latency.

But if the system is distributed, you’ve strewn database parts across multiple hosts, so you lose some of that CPU-to-RAM latency advantage. Or do you?

It’s an interesting design problem with some complex constraints, and it’s the focus of today’s podcast. Our guests are Swapnil Bawaskar, principal engineer at Pivotal; and Jim Bedenbaugh, Advisory Data Engineer at Pivotal.

Swapnil, Jim, and the Datanauts define in-memory databases, describe how they differ from traditional databases, and talk about use cases. They drill into the system architecture for in-memory DBs, including hosts, distributed systems, and failure scenarios. They also discuss practical issues including general operations, metrics, and backup and restoration.

Sponsor: Liquid Technology

Liquid Technology purchases decommissioned IT hardware, provides secure on and off-site data destruction, as well as fully compliant and green e-waste recycling solutions for your organization. Visit liquidtechnology.net/podcast today for a chance to win a $300 Amazon gift card.

Sponsor: Illumio

Illumio’s breakthrough adaptive segmentation technology stops lateral threats inside of any data center or cloud. Illumio works seamlessly between any data center and the public cloud and keeps policies in place as applications move between environments and locations, or auto-scale up/down. Check out their website for details at illumio.com/datanauts.